Anthropic CEO Dario Amodei Calls for Slowing Down AI Development

13 September 2026 · 06:00 · Claude (Anthropic) · claude-sonnet-5

Dario Amodei, CEO of Anthropic, is calling for the breakneck pace of artificial intelligence development to be slowed down. According to him, safety measures are lagging behind technological progress, posing risks for humanity as a whole.

Slowing down AI development has suddenly climbed to the top of the agenda at the world's biggest tech companies. Dario Amodei, CEO of Anthropic, the company behind the chatbot Claude, publicly called this week for the pace of the AI race to be tempered. According to him, the industry owes humanity something: time. Time to let safety measures keep pace with increasingly powerful systems, instead of rushing those systems to market.

What exactly did Amodei say?

In interviews with, among others, The Atlantic, PBS, and CNN, Amodei emphasized that the current pace of artificial intelligence development is out of proportion to the speed at which safety research is progressing. He stated that companies, including his own Anthropic, are under enormous competitive pressure to keep launching more powerful models, while the question of whether those models are actually safe enough is not being adequately answered. When CNN anchor Anderson Cooper asked whether he truly believes AI could harm humanity, Amodei gave a nuanced but troubling answer: the risk is real enough to be taken seriously, even if it is not a certainty.

Why is this call coming now?

The timing of Amodei's remarks is no coincidence. In recent months there has been a series of incidents in which AI systems behaved unpredictably, which, according to RTL Nieuws, shows that the AI race risks spiraling seriously out of control. In addition, a leaked BBC report revealed that former Anthropic employees said they were "genuinely afraid" for humanity's future if the current pace continues. This combination of internal doubts and external incidents appears to be the direct trigger for Amodei to now take a public stand.

Reactions from the AI industry

Notably, Amodei does not appear to be alone in his concerns. According to Politico and the Wall Street Journal, other major players in the sector also acknowledge the need to pump the brakes. Sam Altman of OpenAI admitted in a recent interview that it is "absolutely possible" that AI could eventually slip beyond human control, though he simultaneously stressed that OpenAI is working on robust safeguards. The fact that both Anthropic and OpenAI, two direct competitors in the development of advanced language models, are voicing roughly the same concerns lends extra weight to the discussion. It raises the question of whether competitive pressure can be resolved through self-regulation alone, or whether external, legal frameworks are needed.

What does this mean for users and businesses?

For businesses and consumers who use AI applications such as chatbots, writing assistants, and image generators on a daily basis, a slowdown in development may be barely noticeable in the short term. In the longer term, however, it could mean that new, more powerful models are tested more carefully before becoming widely available. That is good news for anyone concerned about reliability and safety, but it may also mean that some promised breakthroughs take longer to arrive. Anyone wanting to understand how AI has developed over the years can dive into the history of artificial intelligence, which shows how periods of rapid progress have often been followed by moments of reflection.

A turning point for the AI industry?

Amodei's remarks may mark a turning point in the public debate around AI. Where recent years have mostly emphasized who could launch the most powerful models the fastest, there now seems to be cautious room emerging for a conversation about responsibility and pace. Whether this actually leads to concrete agreements between the major players, such as OpenAI, Google, Meta, and Anthropic itself, remains to be seen. What is clear is that the call for responsible AI development is growing louder, including from within the industry itself.

The coming months will show whether words are actually translated into action. Stay informed via more AI news and dig deeper into the subject through our knowledge base for background information on the risks and opportunities of artificial intelligence.

TechCrunch

Source: TechCrunch